Mutation of BCL-6 gene in normal B cells by the process of somatic hypermutation of Ig genes

Immunoglobulin (Ig) genes are hypermutated in B lymphocytes that are the precursors to memory B cells. The mutations are linked to transcription initiation, but non-Ig promoters are permissible for the mutation process; thus, other genes expressed in mutating B cells may also be subject to somatic hypermutation. Significant mutations were not observed in c-MYC, S14, or alpha-fetoprotein (AFP) genes, but BCL-6 was highly mutated in a large proportion of memory B cells of normal individuals. The mutation pattern was similar to that of Ig genes.

Pip, a novel IRF family member, is a lymphoid-specific, PU.1-dependent transcriptional activator

The immunoglobulin light-chain gene enhancers E kappa 3', E lambda 2-4, and E lambda 3-1 contain a conserved cell type-specific composite element essential for their activities. This element binds a B cell-specific heterodimeric protein complex that consists of the Ets family member PU.1 and a second factor (NF-EM5), whose participation in the formation of the complex is dependent on the presence of DNA-bound PU.1. In this report we describe the cloning and characterization of Pip (PU.1 interaction partner), a lymphoid-specific protein that is most likely NF-EM5. As expected, the Pip protein binds the composite element only in the presence of PU.1; furthermore, the formation of this ternary complex is critically dependent on phosphorylation of PU.1 at serine-148. The Pip gene is expressed specifically in lymphoid tissues in both B- and T-cell lines. When coexpressed in NIH-3T3 cells, Pip and PU.1 function as mutually dependent transcription activators of the composite element. The amino-terminal DNA-binding domain of Pip exhibits a high degree of homology to the DNA-binding domains of members of the interferon regulatory factor (IRF) family, which includes IRF-1, IRF-2, ICSBP, and ISGF3 gamma.

Somatic hypermutation of immunoglobulin genes is linked to transcription initiation

To identify DNA sequences that target the somatic hypermutation process, the immunoglobulin gene promoter located upstream of the variable (V) region was duplicated upstream of the constant (C) region of a kappa transgene. Normally, kappa genes are somatically mutated only in the VJ region, but not in the C region. In B cell hybridomas from mice with this kappa transgene (P5'C), both the VJ region and the C region, but not the region between them, were mutated at similar frequencies, suggesting that the mutation mechanism is related to transcription. The downstream promoter was not occluded by transcripts from the upstream promoter. In fact, the levels of transcripts originating from the two promoters were similar, supporting a mutation model based on initiation of transcripts. Several "hot-spots" of somatic mutation were noted, further demonstrating that this transgene has the hallmarks of somatic mutation of endogenous immunoglobulin genes. A model linking somatic mutation to transcription-coupled DNA repair is proposed.

Pip, a lymphoid-restricted IRF, contains a regulatory domain that is important for autoinhibition and ternary complex formation with the Ets factor PU.1

Pip is a lymphoid-restricted IRF transcription factor that is recruited to composite elements within immunoglobulin light-chain gene enhancers through a specific interaction with the Ets factor PU.1. We have examined the transcriptional regulatory properties of Pip as well as the requirements for its interaction with PU.1 and DNA to form a ternary complex. We demonstrate that Pip is a dichotomous regulator; it specifically stimulates transcription in conjunction with PU.1, but represses alpha/beta-interferon-inducible transcription in the absence of PU.1. Thus, during B-cell activation and differentiation, Pip may function both as an activator to promote B cell-specific gene expression and as a repressor to inhibit the antiproliferative effects of alpha/beta-interferons. Mutational analysis of Pip reveals a carboxy-terminal segment that is important for autoinhibition of DNA binding and ternary complex formation. A domain of Pip containing this segment confers autoinhibition and PU.1-dependent binding activity to the DNA-binding domain of the related IRF family member, p48. On the basis of these and other data we propose a model for PU.1/Pip ternary complex formation.

Expression of a microinjected immunoglobulin gene in the spleen of transgenic mice

Transgenic mice were produced by microinjection of a rearranged, functional immunoglobulin kappa gene into fertilized mouse eggs and implantation of the microinjected embryos into foster mothers. Mice that integrated the injected gene were mated and the DNA, RNA and serum kappa chains of their offspring were analysed. The data from offspring of three different transgenic mice indicate that the microinjected gene is expressed in the spleen, but not the liver of mice which inherited the injected gene.

A strain-specific modifier on mouse chromosome 4 controls the methylation of independent transgene loci

A transgene, pHRD, is highly methylated in 12 independent mouse lines when in a C57BL/6 strain background, but becomes progressively less methylated when bred into a DBA/2 background. Transgenes inherited from the mother are generally more methylated; however, this parental effect disappears following continued breeding into the nonmethylating strain. Mapping experiments using BXD recombinant inbred mice as well as other inbred strains indicate that a single strain-specific modifier (Ssm-1) linked to, but distinct from, Fv-1 is responsible for the strain effect. In addition to the methylated and unmethylated transgenic phenotypes, certain mice exhibit a partial methylation pattern that is a consequence of an unusual cellular mosaicism. The pHRD transgene, containing target sequences for the V(D)J recombinase, undergoes site-specific recombination only in lymphoid tissues. This V-J joining is restricted primarily to unmethylated transgene copies.

Allelic exclusion and control of endogenous immunoglobulin gene rearrangement in kappa transgenic mice

Hybridomas were produced from spleen cells of kappa transgenic mice to investigate expression of the transgenic kappa gene, its effect on allelic exclusion and its effect on the control of light-chain gene rearrangement and expression. Our results show that the transgene is expressed normally and that the production of a complete immunoglobulin molecule turns off light-chain gene rearrangement.

Somatic mutation of immunoglobulin light-chain variable-region genes

A single germline immunoglobulin kappa-variable-region gene, VK167, is rearranged and expressed in two myelomas, MOPC167 and MOPC511. Only this single germline gene displays close homology to the expressed genes. Neither of the rearranged, functional genes, however, has a nucleotide sequence that is identical to the germline VK167 gene. Both active genes display several single-base-pair mutations with respect to the germline sequence. The nucleotide sequence data predict the alteration of a restriction-enzyme-recognition site within the VK167 gene between germline cells and cells producing the MOPC167 light-chain protein. Based on this restriction-site alteration, Southern blot analysis proves unambiguously that no gene present in the germline BALB/c mouse genome contains the exact VK167 nucleotide sequence found in cells committed to MOPC167 antibody production. Instead the alterations found in the expressed MOPC167 and MOPC511 V-region genes have apparently arisen by a process of somatic mutation during cellular differentiation. Since nucleotide alterations are found in framework and hypervariable portions of the variable region, the mechanism of somatic mutation is not limited to hypervariable sequences. In addition, Southern blot hybridization indicates that the observed mutations did not arise by recombinational events, but are single-base-pair substitutions. Based on the distribution of mutations that have been found in expressed immunoglobulin variable-region genes, a model that links the introduction of somatic mutations to DNA replication during the V-J joining event is proposed.

AID in somatic hypermutation and class switch recombination

Somatic hypermutation and class-switch-recombination are initiated by the deamination of deoxycytosine in DNA by activation-induced-deaminase, AID. Recently, there has been much research into how AID targets double-stranded DNA in sub-regions of Ig genes, the involvement of co-factors and posttranslational modifications in this process, the co-option of DNA 'repair' mechanisms and AID evolution.

Transgenic mice with mu and kappa genes encoding antiphosphorylcholine antibodies

Transgenic mice were produced that carried in their germlines rearranged kappa and/or mu genes with V kappa and VH regions from the myeloma MOPC-167 kappa and H genes, which encode anti-PC antibody. The mu genes contain either a complete gene, including the membrane terminus (mu genes), or genes in which this terminus is deleted and only the secreted terminus remains (mu delta mem genes). The mu gene without membrane terminus is expressed at as high a level as the mu gene with the complete 3' end, suggesting that this terminus is not required for chromatin activation of the mu locus or for stability of the mRNA. The transgenes are expressed only in lymphoid organs. In contrast to our previous studies with MOPC-21 kappa transgenic mice, the mu transgene is transcribed in T lymphocytes as well as B lymphocytes. Thymocytes from mu and kappa mu transgenic mice display elevated levels of M-167 mu RNA and do not show elevated levels of kappa RNA, even though higher than normal levels of M-167 kappa RNA are detected in the spleen of these mice. Approximately 60% of thymocytes of mu transgenic mice produce cytoplasmic mu protein. However, despite a large amount of mu RNA of the membrane form, mu protein cannot be detected on the surface of T cells, perhaps because it cannot associate with T cell receptor alpha or beta chains. Mice with the complete mu transgene produce not only the mu transgenic mRNA but also considerably increased amounts of kappa RNA encoded by endogenous MOPC-167 like kappa genes. This suggests that B cells are selected by antigen (PC) if they coexpress the mu transgene and appropriate anti-PC endogenous kappa genes. Mice with the mu delta mem gene, however, do not express detectable levels of the endogenous MOPC-167 kappa mRNA. Like the complete mu transgene, the M-167 kappa transgene also causes amplification of endogenous MOPC-167 related immunoglobulins; mice with the kappa transgene have increased amounts of endogenous MOPC-167-like mu or alpha or gamma in the spleen, all of the secreted form. Implications for the regulation of immunoglobulin gene expression and B cell triggering are discussed.

Feedback inhibition of immunoglobulin gene rearrangement by membrane mu, but not by secreted mu heavy chains

Previous work (6-10) has shown that allelic exclusion of Ig gene expression is controlled by functionally rearranged mu and kappa genes. This report deals with the comparison of membrane mu (micron) and secreted mu (microsecond) in promoting such feedback inhibition. Splenic B cell hybridomas were analyzed from transgenic mice harboring a rearranged kappa gene alone or in combination with either an intact rearranged mu gene or a truncated version of the mu gene. The intact mu gene is capable of producing both membrane and secreted forms of the protein, while the truncated version can only encode the secreted form. The role of the microsecond was also tested in pre-B cell lines. Analysis of the extent of endogenous Ig gene rearrangement revealed that (a) the production of micron together with kappa can terminate Ig gene rearrangement; (b) microsecond with kappa does not have this feedback effect; (c) microsecond may interfere with the effect of micron and kappa; and (d) the feedback shown here probably represents a complete shutoff of the specific recombinase by micron + kappa; the data do not address the question of mu alone affecting the accessibility of H genes for rearrangement.

Activation-induced cytidine deaminase (AID) can target both DNA strands when the DNA is supercoiled

The activation-induced cytidine deaminase (AID) is required for somatic hypermutation (SHM) and class-switch recombination of Ig genes. It has been shown that in vitro, AID protein deaminates C in single-stranded DNA or the coding-strand DNA that is being transcribed but not in double-stranded DNA. However, in vivo, both DNA strands are mutated equally during SHM. We show that AID efficiently deaminates C on both DNA strands of a supercoiled plasmid, acting preferentially on SHM hotspot motifs. However, this DNA is not targeted by AID when it is relaxed after treatment with topoisomerase I, and thus, supercoiling plays a crucial role for AID targeting to this DNA. Most of the mutations are in negatively supercoiled regions, suggesting a mechanism of AID targeting in vivo. During transcription the DNA sequences upstream of the elongating RNA polymerase are negatively supercoiled, and this transient change in DNA topology may allow AID to access both DNA strands.

High expression of cloned immunoglobulin kappa gene in transgenic mice is restricted to B lymphocytes

Immunoglobulin genes are normally expressed only in cells of the B lymphocyte lineage after a variable (V) and constant (C) gene rearrangement has occurred. To study the control of immunoglobulin gene expression in a defined situation, we have produced transgenic mice by microinjecting a rearranged mouse immunoglobulin kappa gene (designated pB1-14) into fertilized mouse eggs. We present here the analysis of six different kappa-transgenic mouse lines. All the transgenic mice express the microinjected kappa gene in a completely tissue-specific fashion. Transcripts from pB1-14 are found at a high level in the spleen, but are undetectable in nonlymphoid tissues of testis, liver, kidney, heart, muscle, brain and thyroid gland. In lymphoid cell subpopulations, the level of pB1-14 transcripts is correlated with the relative number of B cells; there is no correlation with the proportion of T lymphocytes. We concluded, therefore, that the microinjected kappa gene contains target sequences for B lymphocyte-specific gene activation signals that override the influence of the integration site.

The E box motif CAGGTG enhances somatic hypermutation without enhancing transcription

The frequency of somatic hypermutations of an Ig kappa transgene with an artificial test insert, RS, is at least 4-fold higher than that of three related transgenes. The four transgenes differ only in the sequence of a 96 bp insert within the variable region. RS is hypermutable over the total 625 nucleotides of the variable/joining region. The RS insert contains two CAGGTG sequences, potential binding sites for basic helix-loop-helix proteins. Changing CAGGTG to AAGGTG reduces the mutability to that of the non-RS transgenes without altering the mutation pattern. The CAGGTG motif enhances somatic hypermutation without enhancing transcription. A DNA probe containing the two CAGGTG sites, but not AAGGTG, binds E47 and gives rise to two specific EMSA bands with nuclear extracts from mutating cells. Possible actions of this enhancer of somatic hypermutation are discussed.

Somatic Hypermutation and Class Switch Recombination in Msh6−/−Ung−/−Double-Knockout Mice

Somatic hypermutation (SHM) and class switch recombination (CSR) are initiated by activation-induced cytosine deaminase (AID). The uracil, and potentially neighboring bases, are processed by error-prone base excision repair and mismatch repair. Deficiencies in Ung, Msh2, or Msh6 affect SHM and CSR. To determine whether Msh2/Msh6 complexes which recognize single-base mismatches and loops were the only mismatch-recognition complexes required for SHM and CSR, we analyzed these processes in Msh6(-/-)Ung(-/-) mice. SHM and CSR were affected in the same degree and fashion as in Msh2(-/-)Ung(-/-) mice; mutations were mostly C,G transitions and CSR was greatly reduced, making Msh2/Msh3 contributions unlikely. Inactivating Ung alone reduced mutations from A and T, suggesting that, depending on the DNA sequence, varying proportions of A,T mutations arise by error-prone long-patch base excision repair. Further, in Msh6(-/-)Ung(-/-) mice the 5' end and the 3' region of Ig genes was spared from mutations as in wild-type mice, confirming that AID does not act in these regions. Finally, because in the absence of both Ung and Msh6, transition mutations from C and G likely are "footprints" of AID, the data show that the activity of AID is restricted drastically in vivo compared with AID in cell-free assays.

The molecular basis of somatic hypermutation of immunoglobulin genes

Somatic hypermutation amplifies the variable region repertoire of immunoglobulin genes. Recent experimental evidence has thrown light on various molecular models of somatic hypermutation. A link between somatic hypermutation and transcription coupled DNA repair is shaping up.

A novel enhancer in the immunoglobulin lambda locus is duplicated and functionally independent of NF kappa B

As a first step toward defining the elements necessary for lambda immunoglobulin gene regulation, DNase I hypersensitive sites were mapped in the mouse lambda locus. A hypersensitive site found 15.5 kb downstream of C lambda 4 was present in all the B-cell but not in the T-cell lines tested. This site coincided with a strong B-cell-specific transcriptional enhancer (E lambda 2-4). This novel enhancer is active in myeloma cells, regardless of the status of endogenous lambda genes, but is inactive in a T-cell line and in fibroblasts. The enhancer E lambda 2-4 functions in the absence of the transcription factor NF kappa B, which is necessary for kappa enhancer function. No evidence could be found for NF kappa B binding by this element. Rearrangement of V lambda 2 to JC lambda 3 or JC lambda genes deletes E lambda 2-4; however, a second strong enhancer was found 35 kb downstream of C lambda 1, which cannot be eliminated by lambda gene rearrangements. The second lambda enhancer (E lambda 3-1) is 90% homologous to the E lambda 2-4 sequence in the region determined to comprise the active enhancer and likewise lacks the consensus binding site for NF kappa B. The data support a model for the independent activation of kappa and lambda gene expression based on locus-specific regulation at the enhancer level.

Rearranged and germline immunoglobulin kappa genes: different states of DNase I sensitivity of constant kappa genes in immunocompetent and nonimmune cells

The rearrangement of a variable (V) and a constant (C) gene appears to be a necessary prerequisite for immunoglobulin gene expression. Multiple different rearranged kappa genes were found in several mouse myelomas, although these cells produce only one type of kappa chain [Wilson, R., Miller, J., & Storb, U. (1979) Biochemistry 18, 5013--5021]. It is therefore of interest to understand how only one allele within a lymphoid cell becomes expressed, while the other allele remains nonfunctional ("allelic exclusion"). We have studied the chromatin conformation of kappa genes by making use of the preferential digestion of potentially active genes by DNase I described, for example, for globin genes [Weintraub, H., & Groudine, M. (1976) Science (Washington, D.C.) 193, 848--856]. The DNase I sensitivity of kappa genes in myeloma tumors, in a B cell lymphoma, and in liver was determined by hybridization with DNA on Southern blots. It was found that rearranged C kappa genes are DNase I sensitive in myelomas in which several kappa genes are rearranged, regardless of whether the rearranged genes code for the kappa chains synthesized by the cell. Furthermore, the C kappa gene in germline configuration is also DNase I sensitive in a B cell lymphoma; i.e., it is in the same chromatin state as the rearranged C kappa gene which probably codes for the kappa chains produced by the cell. The altered chromatin state appears to be localized: V kappa genes in germline context are not DNase I sensitive in myeloma or B lymphoma cells while C kappa genes present in a kappa gene cluster on the same chromosomes are sensitive. When rearranged, however, the V kappa genes are as sensitive to DNase I as are rearranged C kappa genes. V lambda and C lambda genes are not DNase I sensitive in kappa myelomas. Thus, commitment to kappa gene expression is apparently correlated with a chromatin conformation which confers increased DNase I sensitivity to the DNA in the vicinity of all C kappa genes in the cell. "Allelic exclusion" does not operate on the level of chromatin conformation which can be detected by altered DNase I sensitivity.

Different mismatch repair deficiencies all have the same effects on somatic hypermutation: intact primary mechanism accompanied by secondary modifications

Somatic hypermutation of Ig genes is probably dependent on transcription of the target gene via a mutator factor associated with the RNA polymerase (Storb, U., E.L. Klotz, J. Hackett, Jr., K. Kage, G. Bozek, and T.E. Martin. 1998. J. Exp. Med. 188:689-698). It is also probable that some form of DNA repair is involved in the mutation process. It was shown that the nucleotide excision repair proteins were not required, nor were mismatch repair (MMR) proteins. However, certain changes in mutation patterns and frequency of point mutations were observed in Msh2 (MutS homologue) and Pms2 (MutL homologue) MMR-deficient mice (for review see Kim, N., and U. Storb. 1998. J. Exp. Med. 187:1729-1733). These data were obtained from endogenous immunoglobulin (Ig) genes and were presumably influenced by selection of B cells whose Ig genes had undergone certain mutations. In this study, we have analyzed somatic hypermutation in two MutL types of MMR deficiencies, Pms2 and Mlh1. The mutation target was a nonselectable Ig-kappa gene with an artificial insert in the V region. We found that both Pms2- and Mlh1-deficient mice can somatically hypermutate the Ig test gene at approximately twofold reduced frequencies. Furthermore, highly mutated sequences are almost absent. Together with the finding of genome instability in the germinal center B cells, these observations support the conclusion, previously reached for Msh2 mice, that MMR-deficient B cells undergoing somatic hypermutation have a short life span. Pms2- and Mlh-1-deficient mice also resemble Msh2-deficient mice with respect to preferential targeting of G and C nucleotides. Thus, it appears that the different MMR proteins do not have unique functions with respect to somatic hypermutation. Several intrinsic characteristics of somatic hypermutation remain unaltered in the MMR-deficient mice: a preference for targeting A over T, a strand bias, mutational hot spots, and hypermutability of the artificial insert are all seen in the unselectable Ig gene. This implies that the MMR proteins are not required for and most likely are not involved in the primary step of introducing the mutations. Instead, they are recruited to repair certain somatic point mutations, presumably soon after these are created.

Methylation patterns of immunoglobulin genes in lymphoid cells: correlation of expression and differentiation with undermethylation

Different states of eukaryotic gene expression are often correlated with different levels of methylation of DNA sequences containing structural genes and their flanking regions. To assess the potential role of DNA methylation in the expression of immunoglobulin genes, which require complex rearrangements prior to expression, methylation patterns were examined in cell lines representing different stages of lymphocyte maturation. Methylation of the second cytosine in the sequence 5' C-C-G-G 3' was determined by using Hpa II/Msp I endonuclease digestion. Four CH genes (C mu, C delta, C gamma 2b, and C alpha), C kappa, V kappa, C lambda, and V lambda genes were analyzed. The results lead to the following conclusions: (i) transcribed immunoglobulin genes are undermethylated; (ii) the C gene allelic to an expressed C gene is always also undermethylated; and (iii) all immunoglobulin loci tend to become increasingly undermethylated as B cells mature.

Evolution of mouse immunoglobulin lambda genes

The mouse has four C lambda and two V lambda genes. We have isolated Charon 4A clones that contain all six lambda genes from a BALB/c germ-line library. We present here the DNA sequences of the C lambda 2, C lambda 3, and C lambda 4 genes and also correct what are apparently errors in previous reports of C lambda 1 protein and DNA sequences. In addition, we have analyzed cloned DNAs by restriction mapping and electron microscopy to determine the relationships among the various lambda genes. By heteroduplex analysis, two gene clusters containing JC lambda 3--JC lambda 1 and JC lambda 2--JC lambda 4 show homology extending from the J regions 5' of C lambda 3/C lambda 2 to just 3' of C lambda 1/C lambda 4. Other than the region between the genes, very little homology exists in the C lambda flanking regions. In contrast, V lambda 1 and V lambda 2 genes show considerable homology extending into the 5' flanking regions. Large inverted repeats are found in the 5' flanking regions of V lambda 1 and C lambda 3, as well as in the 3' flanking regions of both C lambda gene clusters. DNA sequence divergences between the C lambda genes indicate that an ancestral JC lambda x--JC lambda g gene cluster arose at about the time of the first mammals by duplication of a primordial JC lambda gene. The data further suggest that the JC lambda x--JC lambda gene cluster duplicated after the speciation of mouse and man and subsequently diverged into the present day JC lambda 3--JC lambda 1 and JC lambda 2--JC lambda 4 gene clusters. C lambda 4, a pseudogene, became inactive at about the time of duplication of the ancestral JC lambda x--JC lambda y cluster. Comparison of DNA sequence divergence between the V lambda 1 and V lambda 2 genes demonstrates an anomaly. The percentage of amino acid replacement changes is approximately the same for V lambda 1/V lambda 2 as for C lambda 3/C lambda 2, implying that the ancestral V lambda gene was duplicated at the same time, and possibly together with, the JC lambda x--JC lambda y cluster. However, there are fewer silent changes than amino acid replacement changes between the V lambda 1/V lambda 2 genes, suggesting either that a selective pressure acted on the silent sites or that V lambda genes have only recently been duplicated. We also consider the possibility of a gene conversion event subsequent ot a more ancient duplication.

Somatic hypermutation of an immunoglobulin transgene in kappa transgenic mice

Initial studies of somatically acquired mutations in immunoglobulin V regions from hybridomas and myelomas that are not derived from joining aberrations, suggested a controlled and specific hypermutation process, because spontaneous mutation rates observed for other genes are extremely low. Some evidence for the idea that mutations are introduced during V-gene rearrangement came from the clustering of mutations at the joining sites, from the absence of mutations in unrearranged V genes and from the low level of mutations in only partially (D-J) rearranged nonproductive heavy-chain alleles. Another model in which mutations accumulate with each cell division, rather than being introduced all at once, was supported by the finding that immunoglobulin genes of hybridomas derived from a single mouse frequently had several mutations in common, and so might be derived from the same precursor cell whose daughters then accumulated additional mutations. But the common mutations in some cases could be due to as yet unidentified related germline genes, or could represent the effect of antigen selection for certain amino acids. To try to detect hypermutation in the absence of V-gene rearrangement, we isolated B lymphocytes with endogenous heavy-chain gene mutations from transgenic mice carrying pre-rearranged kappa-transgenes. We found that these kappa-transgenes were also somatically mutated. This and other observations indicated that: ongoing rearrangement is not required for mutation; there are signals for hypermutation in the transgenes; the mutations are found only in the variable region, so the constant region may not be a target; different transgene insertion sites are compatible with hypermutations and more than one transgene is expressed in the same cell.

Effects of sequence and structure on the hypermutability of immunoglobulin genes

Somatic hypermutation (SHM) is investigated in related immunoglobulin transgenes that differ in a short artificial sequence designed to vary the content of hotspot motifs and the potential to form RNA or DNA secondary structures. Mutability depends on hotspots, not secondary structure. Hotspot motifs predict about 50% of the mutations; the rest are in neutral and coldspots. Clusters of mutations and the sequential addition of mutations found in cell pedigrees suggest epigenetic attributes of SHM. Sometime in SHM, an essential factor seems to become limiting. Particular error-prone DNA polymerases appear to create mutations in hotspots on the top and bottom DNA strands throughout the target and the SHM process. One transgene is superhypermutable in all regions, suggesting the presence of a cis-element that enhances SHM.